Method, apparatus, and recording medium for encoding/decoding image by using geometric partitioning

Abstract

Disclosed herein are a method, an apparatus, and a storage medium for image encoding/decoding using geometric partitioning. Partitioning of a block may be determined using directionality in a block. A straight line is detected in the block, some of partitioning methods available for the block are excluded depending on the type of the detected straight line, and only the remaining partitioning methods are limited to a target of selection for encoding and/or decoding on the target block. Further, the mode of a geometric partitioning mode may be determined using directionality in the block. A straight line is detected in the block, some of available modes of a geometric partitioning mode are excluded depending on the type of the detected straight line, and the remaining modes are limited to the target of selection for encoding and/or decoding on the target block.

1 . An encoding method, comprising:
 determining a prediction mode required for prediction for a target block; and   generating information about an encoded target block by performing encoding on the target block using the prediction mode.   
     
     
         2 . The encoding method of  claim 1 , wherein:
 the target block is encoded using a Geometric Partitioning Mode (GPM), and   a mode of the GPM is determined using directionality in the target block.   
     
     
         3 . The encoding method of  claim 1 , wherein a partitioning mode of the target block is determined using directionality in the target block. 
     
     
         4 . The encoding method of  claim 3 , wherein the directionality is derived using a Hough transform. 
     
     
         5 . The encoding method of  claim 4 , wherein the Hough transform is applied to an edge map of the target block. 
     
     
         6 . The encoding method of  claim 5 , wherein an edge in the edge map is detected using one or more of a Sobel operation, a Laplacian operation, and a Canny operation. 
     
     
         7 . The encoding method of  claim 6 , wherein, when a magnitude of a value of the edge is less than or equal to a threshold value, partitioning of the target block is not applied. 
     
     
         8 . The encoding method of  claim 4 , wherein a cumulative frequency threshold value of the Hough transform is determined based on a size of the target block. 
     
     
         9 . The encoding method of  claim 4 , wherein the Hough transform is used to detect a straight line in a horizontal direction, a vertical direction, a 45° diagonal direction, and a 135° diagonal direction. 
     
     
         10 . The encoding method of  claim 4 , wherein the directionality is determined by a straight line detected in the target block. 
     
     
         11 . The encoding method of  claim 1 , wherein a part of available partitioning modes are excluded from encoding on the target block based on a straight line detected in the target block. 
     
     
         12 . The encoding method of  claim 11 , wherein, when no straight line is detected in the target block, none of the partitioning modes are used. 
     
     
         13 . The encoding method of  claim 11 , wherein, when neither of a horizontal straight line and a vertical straight line are detected in the target block, a horizontal partitioning method and a vertical partitioning method are not applied to the target block. 
     
     
         14 . The encoding method of  claim 11 , wherein, when a vertical straight line is not detected in the target block, a vertical partitioning method is not applied to the target block. 
     
     
         15 . The encoding method of  claim 14 , wherein the vertical partitioning method comprises a vertical binary tree (BT) partitioning method and a vertical ternary tree (TT) partitioning method. 
     
     
         16 . The encoding method of  claim 1 , wherein the partitioning mode of the target block is determined using texture property of the target block. 
     
     
         17 . The encoding method of  claim 16 , wherein a part of available partitioning modes are excluded from encoding on the target block based on the texture property of the target block. 
     
     
         18 . The encoding method of  claim 16 , wherein the texture property includes one or more of an edge, a variance, and an average. 
     
     
         19 . A decoding method, comprising:
 determining a prediction mode required for prediction for a target block using a bitstream; and   performing decoding on the target block using the prediction mode and information about an encoded target block.   
     
     
         20 . A non-transitory computer-readable storage medium storing a bitstream for image decoding, the bitstream comprising information about an encoded target block, wherein:
 a prediction mode required for prediction for the target block is determined using the bitstream, and   decoding is performed on the target block using the prediction mode and the information about the encoded target block.